Cancer Therapy Vol 2, 279-290, 2004

Laboratory of Molecular Oncology, Department of Science and Technology, Quilmes National University, Bernal, Buenos Aires, Argentina

__________________________________________________________________________________

*Correspondence: Dr. Daniel E. Gomez, R. Sáenz Peña 180, Bernal B1876BXD Buenos Aires, Argentina; Phone: +54 11 4365-7100 extension 171; Fax: +54 11 4365-7132; e-mail: degomez@unq.edu.ar

Key words: Desmopressin, coagulation, fibrinolysis, surgery, breast cancer, metastasis

Abbreviations: Desmopressin (DDAVP); extracellular matrix (ECM); tissue-type plasminogen activator (tPA)

Peptide hormones released at the neurohypophysis are derived from neurophysins, and display a wide spectrum of biological properties. Oxytocin induces milk ejection and contraction of the uterus, while vasopressin is mainly involved in water balance, causing antidiuresis and increase in blood pressure (North 1987). As shown in Figure 1, the peptide sequence of vasopressin includes 9 aminoacid residues, having a disulfide bridge between positions 1 and 6.

Desmopressin (DDAVP, 1-deamino-8-D-arginine vasopressin) is a synthetic analog of vassopresin described for the first time during the sixties (Zaoral et al, 1967). With homocystein deamination in sequence position 1 the antidiuretic effect is prolonged and the substitution of D-arginine for L-arginine in position 8 decreases the pressor effect of the molecule (Figure 1). After an endovenous dose of 2-20 m g, DDAVP has a plasma half-life between 50 and 160 min. When administered by intranasal route, the half-life is about of 90 min. Although DDAVP is absorbed orally, the doses needed to reach an antidiuretic effect are much higher than the ones needed using the endovenous route. Metabolization of DDAVP is carried out in liver and kidney but slower than vasopressin. Approximately 60% of the compound is released by the kidney without metabolization (Richardson and Robinson 1985).

In contrast to vasopressin, which binds to different cell membrane receptors (V1a, V1b, V2 and V3), DDAVP is a selective agonist for the V2 receptor. This vasopressin

Figure 1. Chemical structure of the nonapeptide hormone vasopressin. The synthetic analog DDAVP differs from the natural hormone by deamination of homocystein at position 1 and D-arginine substitution at position 8 (arrowheads).

receptor subtype is expressed in the kidney collecting duct and mediates the antidiuretic effect of the hormone (Kaufmann et al, 2003a). The V2 receptor is also expressed in endothelial cells (Kaufman et al, 2003b), mediating most of the non-renal effects of DDAVP. Interestingly, the presence of vasopressin receptors was reported in transformed epithelial cells, and also documented in several tumor variants, including breast and lung cancer (North 2000). In addition, neuropeptide receptor expression was detected in different human tumor cell lines (Petit et al 2001).

In the kidney collecting duct, DDAVP activates V2 receptors and causes water retention by inducing the translocation of the water channel aquaporin-2 from intracellular stores to the apical plasma membrane, an example of cAMP-mediated exocytosis (Topal et al, 2003).

The von Willebrand factor is a large glycoprotein playing a role in primary haemostasis, by mediating adhesion of platelets to the subendothelium. It also functions as a carrier protein for coagulation factor VIII, protecting it from proteolytic degradation. The von Willebrand factor is synthesized as a precursor protein in endothelial cells and megakaryocytes. This precursor undergoes dimerization, glycosylation, proteolytic cleavage into von Willebrand factor, and propeptide assembly of the dimmers into large multimers (500-15,000 kDa). Multimerized von Willebrand factor, together with equimolar amounts of propeptide, is stored in specialized secretory granules called Weibel-Palade bodies (Kaufmann et al, 2003a). DDAVP-induced secretion of von Willebrand factor results from V2 receptor-mediated, cAMP-dependent exocytosis from Weibel-Palade bodies.

Some patients treated with repetitive doses of DDAVP during short periods present a progressive decrease in the response of coagulation factor VIII and of von Willebrand factor. Probably, this fact is related with a negative feed-back in the receptors of endothelial cells. Another interesting fact is that taquifilaxia does not occur regarding the release of catecholamines, tPA or urokinase (Vicente et al, 1991).

DDAVP induces an increase in plasma levels of coagulation factor VIII, a cofactor of activated coagulation factor IX, responsible for the activation of factor X of the intrinsic coagulation pathway, leading to the formation of a fibrin clot.

The effect of DDAVP on circulating levels of coagulation factor VIII remains poorly understood. The plasma level of any substance results from the balance between production and removal. Thus, DDAVP could induce factor VIII release from its producing cells. Alternately, factor VIII could be protected from proteolytic degradation, by DDAVP-induced increase in plasma von Willebrand factor, as explained above.

The role of DDAVP in fibrinolysis was one of the first effects described. The profibrinolytic activity of DDAVP is due to an increase in tPA, a proteolytic enzyme that converts plasminogen to plasmin and thus initiates fibrin degradation. The vascular endothelium is thought to be the main source of plasma tPA. In cultured endothelial cells, tPA is expressed at low levels. Its synthesis is up-regulated, usually at the transcriptional level, in response to fluid shear stress, thrombin, histamine, retinoic acid, vascular endothelial growth factor and sodium butyrate. In addition, there is both in vivo and in vitro evidence that tPA is acutely released from preformed stores. A rapid increase in plasma tPA levels is observed in response to DDAVP, as well as beta-adrenergic agents administered systemically (Wall et al, 1998).

Co-localization of von Willebrand factor and tPA in the same compartment could account for the coordinate effect of DDAVP on the plasma level of the two proteins. The identification of a storage compartment for tPA, distinct from Weibel-Palade bodies, remains unexplained (Emeiss et al, 1997).

DDAVP is known to have vasodilator properties, as shown by an increase in heart rate and a decrease in systolic and diastolic blood pressure, as well as facial flushing (Derkx et al, 1983). Perfusion studies have demonstrated that vasopressin and DDAVP exert a direct vasodilator effect after intraarterial administration by a mechanism dependent of nitric oxide (Hayoz et al, 1997). These observations suggest a direct activation of endothelial nitric oxide synthase in the skeletal muscle vasculature, in a V2 receptor-dependent, cAMP-mediated manner.

The adhesion molecule P-selectin is expressed in both endothelial cells and megakaryocytes/platelets, in Weibel-Palade bodies and ·-granules, respectively (McEver et al, 1989). Kanwar et al demonstrated that DDAVP induced a significant but transient increase in P-selectin expression on human umbilical vein endothelial cells, as well as on rat and human platelets. Earlier studies have shown that endothelial cell expression of P-selectin is important for the very early leukocyte-endothelial cell interaction, known as leukocyte rolling, an absolute prerequisite for leukocyte adhesion and migration (Kanwar et al, 1995).

As blood monocytes have been identified as a target for DDAVP, Pereira et al demonstrated that DDAVP enhanced the ability of blood monocytes to bind activated platelets, mainly by increasing the expression of P-selectin sialylated ligands on the monocyte surface (Pereira et al, 2003).

Researchers have shown that intravenous DDAVP (0.3 m g/kg) increased 2-fold the plasma levels of norepinephrine (Grant et al, 1988). Concomitantly, other authors have demonstrated that central and peripheral administration of DDAVP increase locomotor activity in rats in doses that alter brain dopamine neurochemistry. By using different catecholamine manipulating agents, they reported that the central stimulatory action of DDAVP involves granula-mediated dopamine release and subsequent activation of dopamine receptors, and that alpha-adrenoceptors possibly also are involved (Di Michele et al, 1988).

The primary adverse reaction associated with DDAVP is hypotonic hyponatremia. Hyponatremia has been reported in adults treated with DDAVP for Von Willebrand´s disease and hemophilia, and in children and adults treated for enuresis (Shulman et al, 1990). Water intoxication is uncommon when the drug is used with proper precautions. The strong antidiuretic action of DDAVP has some potential problems that are negligible in adults and older children when water intake is restricted. In infants and small children under the age of 18 months, however, DDAVP should be used with caution in order to prevent water intoxication and electrolyte imbalance. Extreme caution should be exercised when the patients receive parenteral fluid substitution (Sutor, 1998).

Other side effects observed in the treatment of bleeding disorders are mild and transient, including facial flushing, transient headache, increased pulse rate and drop in systolic blood pressure. They can be minimized when the dose is not exceeding 0.3 m g/kg body weight, and the infusion lasts at least 20 to 30 minutes.

Registered thrombotic episodes are few. An interesting review analyzed the number of people treated between 1985 and 1988, estimated in approximately 433,000, and the number of published thrombotic episodes was 10. The author concludes that the prothrombotic risk was of 0.0001% (Rodheghiero et al, 1991).

DDAVP was primarily used for its antidiuretic properties in the treatment of conditions such as central diabetes insipidus and enuresis. In 1977 DDAVP was used for the first time to treat patients with hemophilia A and von Villebrand disease, the most frequent congenital bleeding disorders. The clinical indications for DDAVP quickly expanded beyond these diseases. We will analyze its clinical utility, in order to provide evidence about the widespread use and good tolerance.

Diabetes insipidus is an uncommon condition characterized by polyuria and polydipsia. The symptoms and biochemical changes of this condition result from either a lack of the antidiuretic hormone vasopressin, or renal insensitivity to its effect. Failure to produce or release the hormone may result from cranial pathology. The renal insensitivity to vasopressin that occurs in patients with nephrogenic diabetes insipidus may be caused by genetic factors, drugs (especially lithium) or specific disease processes. Patients may compensate for polyuria and nocturia by excessive water intake but show marked decreases in urine specific gravity and osmolality. Patients with severe and uncompensated symptoms develop dehydration, neurological symptoms and encephalopaty.

Vasopressin "replacement" with DDAVP is the treatment of choice in patients with cranial diabetes insipidus, although extreme caution is required when treating babies or small children because of the danger of fluid overload. The treatment of nephrogenic diabetes insipidus is difficult and typically involves therapy with a diuretic such as chlorothiazide, as well as indomethacin. These agents enhance urine osmolality by their renal effect on solute and water handling (Cheetham et al, 2002).

Von Willebrand disease is an autosomal dominantly inherited hemorrhagic disease caused by a deficiency in von Willebrand factor. Most patients have a mild disease that may go undiagnosed until trauma or surgery. Symptomatic individuals manifest easy bruisability and mucosal surface bleeding. The goals of therapy consist of correcting the deficiencies in von Willebrand factor protein activity to above 50% of normal and coagulation factor VIII activity to levels appropriate for the clinical situation. DDAVP (0.3 m g/kg in endovenous saline infusion or 150 m g intranasally for adults) is the recommended treatment for type 1 von Willebrand disease, eliminating potential exposure to blood-borne pathogens that replacement therapies may contain. DDAVP administration should be avoided in most individuals with type 2B variant of the disease (Mannucci, 1997).

Hemophilias are sex-linked recessive disorders. Hemophilia A is caused by a deficiency of coagulation factor VIII and hemophilia B is caused by a deficiency of factor IX. A deficiency of either of these two intrinsic coagulation pathway components result in inefficient and inadequate generation of thrombin, which cannot be circumvented or supplemented by a normal extrinsic pathway because of the strong modulatory effects of tissue factor pathway inhibitor. Severe cases are characterized by frequent spontaneous bleeding events in joints (hemarthrosis) and soft tissues, and by profuse hemorrhage with trauma or surgery. DDAVP is useful in patients with mild hemophilia A since an adecuate incremental rise in factor VIII activity can circumvent the use of clotting factor concentrates (De La Fuente et al, 1995).

Regarding congenital trombocitopenias, the use of DDAVP varies according with the pathology. Storage pool disease is an autosomal dominant disorder, whereas platelet storage granules are decreased in number and/or content, presumably because of abnormal granule formation in megakaryocytes. The bleeding diathesis is mild and affects mostly women. With the deficiency in dense granules, platelets aggregate abnormally. Dense-granule storage pool disease is also associated with several other congenital disorders, including oculocutaneous albinism in both the Hemansky-Pudlak and Chédiak-Higashi syndromes.

The Bernard-Soulier syndrome, is an autosomal recessive disorder caused by a deficiency of a platelet membrane glycoprotein complex. As a result, giant platelets appear in the peripheral blood smear. Physiologically, platelets fail to adhere normally to subendothelial connective tissue because of defective binding of Von Willebrand factor. Glanzmann´s thrombasthenia, is an autosomal recessive bleeding disorder characterized by a prolonged bleeding time and platelets that fail to aggregate normally when stimulated. Some patients with storage pool disease are responsive to DDAVP administration (Nieuwenhuis et al, 1988). Furthermore, in the Hermansky-Pudlak syndrome the use of ristocetin and collagen with DDAVP produce a shortening in the bleeding time and improve platelet aggregation (Wijermans et al, 1989). In the rest of the pathology analysed although use of DDAVP has improved bleeding time and aggregation, there are no parameters that allow to foresee whether a patient will be responsive or not to DDAVP infusion.

Acquired Von Willebrand disease is a rare condition and usually occurs as a complication of autoimmune, myeloproliferative, or lymphoproliferative disorders. The acquired disease associated with neuroblastoma is secondary to proteolysis of Von Willebrand factor by tumor-secreted hyaluronidase. Abnormal multimeric composition of Von Willebrand factor is a hallmark of these syndromes. Treatment is similar to that for congenital disease, but responses are unpredictable (Tefferi et al, 1997).

Autoantibody inhibitors occur spontaneously in individuals with previously normal hemostasis. Although approximately 50% of the cases have no obvious underlying etiology, the remainder is associated with autoimmune diseases, lymphoproliferative disorders, idiosyncratic drug associations or pregnancy. Patients typically have massive hemorrhagic disorders. Treatment usually includes replacement therapy, porcine factor VIII concentrate and immunosuppressive therapy with steroids and cytotoxic agents. There are some reports of patients with this pathology satisfactory treated with DDAVP (Cohen et al, 1996).

Platelets function abnormally in patiens with renal failure. The uremic metabolites responsible for this disfunction are uncertain, but certain phenolic compounds that accumulate in uremia may inhibit platelet aggregation. Uremic bleeding is usually mucocutaneous and reflects abnormal platelet or vascular hemostatic functions. DDAVP is usually a good and safe alternative for profilaxis and treatment of hemorrhagic alterations associated with terminal uremia (Lens et al, 1988).

Platelet function is sometimes abnormal in liver disease, but the mechanisms and the extent to which it contributes to bleeding are unclear. DDAVP has been reported to improve the bleeding time in these circumstances (Mannucci et al, 1986).

DDAVP counteracts the effect on hemostasis of some antithrombotic drugs. It shortens the prolonged bleeding time of individuals taking widely used antiplatelet agents, such as aspirin and ticlopidine, and the prolonged bleeding time and activated partial thromboplastin time of patients receiving heparin. It also counteracts the antihemostatic effects of dextran, with no apparent impairment of the antithrombotic properties. Although, more clinical evidence is needed, DDAVP may provide an opportunity to control drug-induced bleeding without stopping treatment and perhaps avoiding recurrence or progression of thrombosis (Butler et al, 1993).

Several investigators have evaluated whether DDAVP was beneficial during surgical operations in which blood loss is large and for which multiple blood transfusions are needed. Open-heart surgery with extracorporeal circulation is the epitome of operations that warrant the adoption of blood-saving measures.

Conflicting results using DDAVP in open-heart surgery were obtained and they might be due to the fact that most studies were of small size and had insufficient statistical power to detect true differences in blood loss. A meta-analysis of 17 randomized, double-bind, placebo-controlled trials, which included 1171 patients undergoing open heart surgery, has attempted to overcome this pitfall. Overall, DDAVP reduced postoperative blood loss by 9%. Although DDAVP had no blood-saving effect when the total blood loss was low, the compound seems to be beneficial in cardiac operations associated with blood loss larger than 1 liter (Cattaneo et al, 1998).

Nocturnal enuresis is a prevalent clinical problem in childhood and adolescence. It is a heterogeneous disorder with various underlying mechanisms, causing a mismatch between the nocturnal bladder capacity and the amount of urine produced during sleep at night, in association with a simultaneous failure of conscious arousal in response to the sensation of bladder fullness. Children with increase nocturnal urine production usually have a good response to DDAVP therapy (Eggert et al, 2001). Patients with a small bladder almost invariably have various types of occult bladder dysfunction, but otherwise have a completely normal circadian rhythm of urine production. These patients generally have a poor response to DDAVP treatment, but would benefit more from combination therapy with enuretic alarm, urotherapy and antimuscarinic agents, in addition to DDAVP.

Nocturia is also common in elderly men and women. The circadian rhythm of arginine vasopressin present in younger individuals, is lost in the elderly. The efficacy of DDAVP treatment (0.1 mg oral at bedtime) in patients of 65 years-old and older with nocturia was investigated and found safe and effective (Kwo, 2003).

Having described the current clinical use, we will describe the process of invasion and metastasis, and then analyze the antitumor properties of DDAVP in preclinical animal models. The known body of literature about the compound will be related with the new evidence showing DDAVP as a potential perioperative adjuvant for cancer surgery.

To form secondary growths, cancer cells at the primary tumor must invade the surrounding tissue, penetrate vessels, and travel to other sites where they arrest and resume growth (Figure 2). Metastasis is the major cause of mortality in cancer patients (Fidler, 1991). Of all the tumor cells that enter into the circulation, only 0.01% will survive to produce secondary tumors. Metastatic capacity depends in part on angiogenesis, a process by which the tumor induces the formation of new blood vessels, beginning with capillary buds and progressing to a vascular network. The new blood vessels within and around the tumor mass provide nutrients for tumor growth and create access to circulation for metastasis (Thorgeirsson et al, 1994).

The invasion process can be classically divided into three sequential steps: adhesion of tumor cells to the basement membrane and extracellular matrix (ECM), disruption of the basement membrane by proteolytic digestion, and migration through the modified basement membrane (Liotta, 1986). A biological similitude between tumor invasion and neovascularization underlines a cooperative function of cancer cells and endothelial cells during tumor progression.

Adhesion of tumor cells to the basement membrane involves specific anchoring glycoproteins of ECM, such as fibronectin, laminin and collagens, which bind to a variety of tumor cell surface receptors. To penetrate ECM, the invading cells must disrupt local segments in the organized structure of the basement membrane, a tightly regulated process involving proteolytic enzymes. Once the tumor cells enter the stroma, they can easily gain access to lymphatic and blood vessels for further dissemination.

Four major classes of proteases are important in the invasion process: serine, aspartyl, cysteinyl, and metal ion-dependent proteases. Many subclasses of metalloproteases have been described, including interstitial collagenase, type IV collagenases, and stromelysin. There is evidence that tumor cells elaborate different types of proteases, which together with proteases expressed by surrounding host cells such as endothelial cells, fibroblasts and inflammatory cells, are capable of degrading the complex network of ECM barries (Thorgeirsson et al, 1994).

Tumor invasion and metastasis require active cell motility, not only for the endothelial cells in the process of angiogenesis but also for the tumor cells. Migration is initiated by pseudopodia, followed by translocation of the entire cell. The locomotion involves assembly and disassembly of cross-linked actin filaments, govern by specific cell signals (Gomez et al, 1999). Once the tumor cells gain access to a blood vessel, it is ready to circulate into the blood and reach distant sites.

It has been long-recognized that although a large number of cancer cells may be released each day from primary tumors, comparatively few metastases develop from these cells. This "metastatic inefficiency" has been well-documented by observations on humans, and several experiments in animal models. A precise estimation of the inefficiency of circulating cancer cells in forming tumors, is obtained from counts of pulmonary colonies in mice after receiving tail-vein injections of metastatic tumor cell suspensions (Weiss et al, 1982). Even with highly aggressive transplantable tumors, efficiencies of less than 0.1% are common. When combined with cancer cell loss and delay associated with intravasation, this constitutes a high degree of operational metastatic inefficiency.

Kinetic studies in mice point to the massive destruction of cancer cells in the microcirculation. As a result of interactions with microvessel walls, it appears that some tumor cells are killed relatively slowly, over minutes or hours by various arms of the inflammatory and/or immunologic response, whereas others are killed rapidly over seconds by mechanical damage (Weiss et al, 1983).

Figure 2. Critical steps in tumor invasion and metastasis. 1) Intravasation from a primary growth. 2) Circulation in blood, whereas cells can be destroyed or protected in a tumor emboli. 3) Extravasation. 4) Secondary growth and angiogenesis.

Following this early evidence, recently Topal et al have demonstrated that aggregated colon cancer cells have a higher metastatic efficiency in the liver compared with non-aggregated cells. Hepatic metastases were observed in 81% of the rats after intraportal injection of aggregates equivalent to 0.5 x 10⁶ cancer cells. A significant lower metastatic efficiency (16%) was found after the injection of the same number of non-aggregated cancer cells (Topal et al, 2003). Similar results were obtained by other authors who found that in contrast with viable single or non-aggregated cells that often fail to form metastases, tumor cell clumps result in a high metastatic efficiency after injection via the portal vein (Panis et al, 1992).

Aggregated cancer cells may remain in large clusters of viable cells, and trapped in venous or arterial branches where they get attached to the endothelial cells. Here they may be able to evade host defense mechanisms and form secondary tumors. On the contrary, non-aggregated cancer cells may be unable to form clusters of viable cells and be challenged with mechanical forces and immune defenses.

Metastatic tumor cells entering into the blood stream interact with components of the haemostatic system. This interaction results in fibrin deposition around tumor cells, determining the formation of microthrombi that increase the efficiency of the metastatic cascade (Constantini et al, 1992). Fibrin deposition may determine an enhanced intravascular tumor cell aggregation and trapping in the target organ, and also protects tumor cells from destruction by host immunity (Gunji et al, 1998). In this regard, we have reported an enhancement of lung colonization by mammary tumor cells administering a synthetic inhibitor of the profibrinolytic enzyme urokinase during the first stages of metastasis formation (Alonso et al, 1996).

Although the metastatic process is highly inefficient, any release of tumor cells into the circulation should be avoided. It has been suggested that surgical manipulation can provoke liberation of viable cancer cells. The presence of cancer cells in the peripheral blood has been confirmed by reverse transcription-polymerase chain reaction in patients undergoing breast cancer surgery (Brown et al, 1995). Similarly, conventional chemotherapy may cause a mobilizing effect on cancer cells. Other authors have reported the recruitment of tumor cells into the peripheral blood after the first courses of primary chemotherapy in patients with breast cancer enrolled in a prospective study. (Sabbatini et al, 2000).

Other authors reported the histological findings in a series of axillary lymph node dissections taken approximately 2 weeks after breast biopsy (Carter et al, 2000). They described the presence of epithelial cells in the subcapsular sinus of draining lymph nodes that may be attributed to mechanical transport of tumor breast epithelium secondary to the previous needle or surgical manipulation. Recently, Moore et al investigated by immunohistochemical staining whether the pattern of sentinel lymph node metastasis in breast cancer is related to tumor manipulation. Interestingly, the data suggested that the frequency of positive nodes is increased after instrumentation of the tumor site (Moore et al, 2004).

Several experimental studies with animal models have confirmed that intrabdominal tumor manipulation was the main factor acting on metastatic dissemination using conventional laparotomy or laparoscopy (Mutter et al, 1999). It has been shown that port site tumor recurrence rates decreased with increased surgical experience in a mouse adenocarcinoma model of laparoscopic splenectomy, suggesting that a poor surgical technique was the main cause of recurrence (Lee et al, 2000). In the same line, interesting results were obtained in an experimental model of breast cancer. Syngeneic mice were inoculated into the mammary fat pad with TA3Ha adenocarcinoma cells and the resulting tumors were surgically excised with a curative intent. Under these conditions, perioperative chemotherapy with doxorubicin reduced local recurrence, axillary metastasis, and lung metastasis, and also improved disease-free survival (Murthy et al, 1996).

We have examined the effects of neuropeptide hormones on our mouse mammary carcinoma model F3II (Alonso et al, 1997). We reported that vasopressin and its synthetic derivative DDAVP can modulate tumor cell growth in vitro and the secretion of urokinase, a profibrinolytic enzyme involved in hematogenous metastasis. In this regard, enhancement of pericellular fibrinolysis may prevent coating of intravascular tumor emboli with fibrin, therefore decreasing the survival of tumor cells in the circulation (Alonso et al, 1996).

The formation of multicellular aggregates of mammary tumor cells in the presence of plasma from control or DDAVP-treated mice was investigated in ex vivo assays. After a short time, control plasma induced a significant aggregation of the tumor cell suspension. Also, a clot was formed in tubes and tumor cell clumps were trapped in a fibrin gel matrix. In contrast, in the presence of plasma from DDAVP-treated mice, most of the mammary tumor cells remained as a single cell suspension (Alonso et al, 1999). DDAVP did not reduce cell viability of tumor cell suspensions at the doses employed. Similarly, semiconfluent monolayers were not affected after continue in vitro culture in the presence of DDAVP.

We have examined the effects of DDAVP on experimental lung colonization of highly metastatic mammary tumor cells in syngeneic Balb/c mice. Coinjection of DDAVP (1-2 m g/kg body weight) at the time of endovenous inoculation of F3II carcinoma cells significantly inhibited the formation of experimental lung metastases. Similar results were obtained with the parental LM3 mammary adenocarcinoma cells. In both cases, the number of lung nodules was reduced up to 70% in DDAVP-treated mice. Inhibition of metastasis was also obtained with administration of DDAVP 24 h after tumor cell inoculation (Alonso et al, 1999).

Interestingly, in vitro pretreatment of tumor cells with comparable concentrations of DDAVP followed by peptide washout did not reduce the incidence of lung colonies, ruling out the possibility that DDAVP was mediating its antimetastatic activity through a direct effect on tumor cells. Extrapulmonary tumor colonies were not found in any of the control mice or mice treated with DDAVP. Our experiments suggested for the first time that adjuvant DDAVP therapy can impair successful implantation of circulating cancer cells.

Considering the antimetastatic effect of DDAVP in animal studies, as well as its well-known hemostatic and profibrinolytic properties, the compound is an excellent candidate for adjuvant therapy both during and immediately after tumor surgery. Therefore, we investigated the effect of DDAVP on lymph node and lung metastasis, using a preclinical mouse mammary carcinoma model of subcutaneous tumor manipulation and surgical excision.

We developed an experimental instrument for the application of controlled pressures on subcutaneous tumors. It consists on a mobile platform that transmits pressure though an axis to a small surface of 6 cm². The platform is loaded with the appropriate weight and the instrument discharges a stable and controlled pressure on the tumor mass. Tumor-bearing mice were anesthetized and subcutaneous tumors subjected to experimental manipulations using pressures of 0.5 kg/cm² during 2 min. To examine the antimetastatic properties of DDAVP, subcutaneous tumors were subjected to 2 or 3 weekly experimental manipulations, followed by surgical excision. DDAVP was administered intravenously 30 min before and 24 h after each manipulation or surgery, at a dose of 2 m g/kg. At the end of the experiment, mice were sacrificed and necropsied (Giron et al, 2002).

Tumor manipulation induced massive dissemination to axillary nodes and increased up to 6-fold the number of metastatic lung nodules. Perioperative treatment with DDAVP dramatically reduced regional metastasis. The incidence of lymph node involvement in manipulated animals was 12% with DDAVP therapy and 87% without treatment (Figure 3). Histopathological analysis of most axillary nodes from DDAVP-treated animals showed sinusal histiocytosis and no evidence of cancer cells. Histiocytic reaction of the regional lymph nodes is considered a strong indicator of antitumor resistance in patients with breast cancer (Loboda et al, 1982). In contrast, axillary nodes from control mice bearing manipulated mammary tumors and administered with the saline vehicle evidenced massive metastasis and lacked sinusal histiocytosis. As shown in Figure 3, metastatic lung nodules were also reduced about 65% in animals treated with DDAVP (Giron et al, 2002). Perioperative DDAVP apperead to be safe at the dosage employed, and antitumor resistance was obtained without overt toxic effects.

Figure 3. Effect of perioperative DDAVP on lymph node and lung metastasis in mice bearing subcutaneous F3II mammary carcinoma subjected to repeated experimental manipulation. Each group represent the combined data of a minimum of 6 animals.

*p<0.05 versus its respective control in manipulation plus saline group. Kruskal-Wallis test.

**p<0.02 versus its respective control in manipulation plus saline group. Chi square test.

The biological effects of DDAVP on both endothelial and tumor cells are complex, and further investigations will determine the precise mechanisms of antitumor action. Nevertheless, the hemostatic effect of DDAVP seems to be pivotal, since it improves and accelerates the postoperative healing process. In this context, local and distant recurrence of breast cancer may be because of the perioperative stimulation of residual cancer cells (Reid et al, 1997). The perioperative period is also characterized by immunosuppression that may predispose to tumor spread (Vallejo et al, 2003).

Perioperative DDAVP may offer the opportunity to modulate the early wound environment and reduce locoregional cancer recurrence rates. Enhanced coagulation after tumor manipulation may contribute to a rapid encapsulation of residual tumor tissue, limiting intravasation of tumor cells. It is known that proangiogenic molecules are locally produced in response to wounding and cancer. Recently, very high local concentrations of angiogenic factors were detected in surgical wound fluid samples from breast cancer patients, suggesting that they may need to be antagonized using perioperative systemic or local therapy (Hormbrey et al, 2003).

Other blood-saving agents have been used during cancer surgery. Administration of perioperative and postoperative tranexamic acid reduced the frequency of wound complications in women with breast cancer undergoing lumpectomy or mastectomy (Oertli et al, 1994). Similarly, intraoperative infusion of the hemostatic agent aprotinin, a nonspecific protease inhibitor, was associated with a significant survival benefit in patients underwent liver resection for colorectal cancer metastasis (Lentschener et al, 1999).

In addition, DDAVP increases intravascular fibrinolysis, helping to dissolve the protective fibrin shield of circulating tumor cells and reducing tumor cell aggregation (Alonso et al, 1999). As mentioned above, fibrin deposition around cancer cells entering into the blood stream ameliorates cell survival and facilitates trapping in the target organ. In accordance, implantation of mammary tumor cells at sites of trauma in an experimental mouse model was inhibited by injection of profibrinolytic agents, such as streptokinase and recombinant tPA (Murthy et al, 1991).

DDAVP effect is exerted in the early stages of metastasis, not only by inducing rapid encapsulation of residual tumor tissue and limiting the formation of intravascular tumor cell emboli, but also altering the interaction of cancer cells with endothelium (Table 1). For instance, DDAVP may modify tumor cell attachment at the target organ by altering P-selectin expression on endothelial cells (Kanwar et al, 1995) or platelets (Wun et al, 1995). DDAVP may also alter hemodynamics of blood flow or induce lysis of tumor cells through the production of nitric oxide from the microvasculature (Hirano, 1997). Furthermore, we cannot exclude direct biological effects of DDAVP on tumor cells during intravasation and formation of metastatic foci. It has been described that breast and small-cell lung cancer cells contain normal genes for all vasopressin receptors and express normal vasopressin V1a and V1b receptor proteins, plus both normal and abnormal forms of the V2 receptor (North, 2000).

DDAVP has been used in patients with diabetes insipidus and in a variety of bleeding disorders. DDAVP is a safe and effective hemostatic agent for use during surgery in patients with hemophilia or von Willebrand disease. Antitumor properties of DDAVP in tumor models were obtained administering endovenous doses close to the ones previously used and proved enhanced antidiuretic or hemostatic effect (0.3-4 m g/kg). These doses have the advantage of being well characterized from a pharmacological point of view (Lethagen, 1994; Mannucci, 1997).

Our preclinical observations strongly suggest the application of DDAVP as a perioperative adjuvant in cancer surgery. The potential dual role of DDAVP in surgical oncology, reducing blood loss and limiting tumor recurrence or metastasis, warrant further investigation. If similar findings are obtained in humans, pharmacologic modulation of hemostasis and fibrinolysis using DDAVP should become a priority in the management of cancer patients undergoing surgery.

Surgical manipulation and tissue trauma enhance the growth and dispersement of many types of malignant cells. However, as wounds develop and healing is complete the surgical site becomes less favorable to tumor implantation (Murthy et al, 1989). Thus, local recurrence found in conjunction with widespread metastatic disease is likely to have been established by perioperative seeding rather than as a late phenomenon (Hormbrey et al, 2003).

Available experimental evidence indicates antitumor effects of DDAVP in breast cancer, and similar benefits in other aggressive solid tumors are expected, such as prostate cancer, ovarian cancer, head and neck cancer, colorectal cancer, esophageal cancer, lung cancer, sarcomas, melanoma and central nervous system tumors. In this regard, a panel of synthetic peptide analogs has been developed in our laboratory in the search for improved efficacy in particular tumor variants.

In the future, we will gain a better understanding of the complex biological events that occur during the perioperative period in cancer patients. Whichever the mechanisms of action involved, the hemostatic and profibrinolytic compound DDAVP appears as a new agent which could be able to act cooperatively with cancer surgery, as well as with other standard therapies, to reduce recurrences and improve survival of patients. It seems that perioperative treatment strategies will be a fruitful area for cancer research in the next years and deserve further clinical investigation.

The authors want to thank Genesica for its useful advice and management of the intellectual property of the findings. This work was supported by the R&D Priority Grant Program from Quilmes National University (53-A048) to D.E.G. and D.F.A. To Guillermo Skilton, in memoriam.

From left to right: Dr. Daniel E. Gomez, Dr. Giselle V. Ripoll, Dr. Santiago Girón, Dr. Daniel F. Alonso